Does the Internet worsen polarisation?

Headline: Mad Max public square on the Internet
Writer: David Brooks
Source: The Straits Times, 21/4/10

Quote1
In 2001, Mr Cass R. Sunstein, ... a professor at the University of Chicago, ... raised the possibility that the Internet may be harming the public square.

Comment1
The conclusion is "the Internet harms the public square". What is the argument?

Quote2
In the mid-20th century, Americans got most of their news through a few big networks and mass-market magazines. People were forced to encounter political viewpoints different from their own. ... Mr Sunstein wondered whether the Internet was undermining all this. The new media, he noted, allow you to personalise your newspapers so you see only the stories that already interest you. You can visit only those websites tht confirm your prejudices. Instead of a public square, we could end up with a collection of information cocoons.

Comment2
We are told there is a change from "force-fed information" to "self-selected information" -- and that this can lead to "information cocoons". Here is the argument, formally presented:

P1: If (self-selected information), then (information cocoons)
P2: (self-selected information)
C1: Hence, (information cocoons)

The argument has the valid Modus Ponens (If P then Q / P / Hence, Q) argument form. P2 is true, as a matter of general knowledge. Is P1 also true?

Quote3
Mr Sunstein ... has done ... work ... about our cognitive biases. We like hearing evidence that confirms our suppositions. We filter out evidence that challenges them. We have a ntural tilt towards polarised views. People are prone to gather in like-minded groups.

Comment3
Right. Here is the link that supports P1.

P3: If (self-selected information), then (similar information) [cognitive bias]
P4: If (similar information), then (information cocoons)
C2: Hence, if (self-selected information), then (information cocoons)
C2 = P1

This has the valid Hypothetical Syllogism (If P then Q / If Q then R // Hence, If P then R) argument form. P3 is what Sunstein has found in his work on cognitive bias. P4 is true by definition. The argument is sound. P1 is true.

We return to the argument in Comment2. The argument form is valid. P2 is true. Now, P1 is also true. The argument (P1 / P2 // hence C1) is sound. Therefore, C1 is true -- we will develop information cocoons.

Quote4
Once in them (like-minded groups), they (people) drive one another to even greater extremes. ... Mr Sunstein's fear was that the Internet might lead to a more ghettoised, polarised and insular electorate.

Comment4
Thus, information cocoons in turn lead to greater polarisation:

If (information cocoons), then (greater polarisation)

Quote5
Yet new research complicates this picture. Mr Matthew Gentzkow and Mr Jesse Shapiro, both of the Unviersity Chicago Booth School of Business, have measured ideological segregation on the Internet. ... But the core finding is that most Internet users do not stay within their communities. Most people spend a lot of time on a few giant sites with politically integrated audiences, like Yahoo! News. But even when they leave these integrated sites, they often go into areas where most visitors are not like themselves. ...

Comment5
This is not a complication. This is a direct refutation of P3. This says that when people self-select information, they do not go for similar information. It says P3 is false. This makes argument (P3 / P4 // hence, C2 = P1) unsound. P1 is not proven true. Hence, argument (P1 / P2 // hence, C1) fails.

Quote6
This does not mean they are not polarised.

Comment6
This quote relates to argument (P3 / P4 // hence C2) in Comment3. The word "this" in Quote6 refers to the denial of:

P3: If (self-selected information), then (similar information)

This denial does not lead to a denial of (information cocoons) and hence polarisation.

Here is the whole argument.

P3: If (self-selected information), then (similar information)
P4: If (similar information), then (information cocoons)
P2: (self-selected information)
C1: Hence, (information cocoons)

P5: (information cocoons) [from C1]
P6: If (information cocoons), then (greater polarisation) [from Comment4]
C3: Hence, (greater polarisation)

We deny P3. To go from (not-P3) to (not-C3) will commit the Fallacy of Denying the Antecedent (If P then Q / not-P // hence, not-Q). So we cannot deny (greater polarisation). At the same time, we also cannot assert (greater polarisation), since we denied P3 in Comment3. Result: We do not know about polarisation.

Quote7
Looking at a site says nothing about how you process it or the character of attention you bring to it. It could be people spend a lot of time at their home sites and then go off on forays looking for things to hate. ...

Comment7
This illustrates why Denying the Antecedent is a fallacy. Denying one cause of a phenomenon does not preclude other causes of the same phenomenon.

Quote8
The study also suggests that if there is increased polarisation (and there is), it is probably not the Internet that is causing it.

Comment8
We have not been presented with any argument supporting the claim that there is increased polarisation. This is just the writer's assertion.

From Quote6 and Comment6, we saw that people not visiting confirmatory sites does not lead to any firm conclusion about polarisation. In Quote7, the writer says people could "go off on forays looking for things to hate".

Given these, I do not see how we can arrive at the conclusion that "it is probably not the Internet that is causing it [increased polarisation]."

END